Malcolm J. Drury

Terrestrial Heat Flow in Canada

Abstract Heat flow has been measured at 214 separate sites in Canada and on the surrounding continental shelves. Most geological provinces show average heat flow that conforms approximately to world averages for areas of similar tectonic age, but the northern Prairies show an anomalously high heat flow that is probably controlled by deep water movement, and the coastal zone of southern British Columbia shows a low heat flow that is believed to be associated with recent subduction of small ocean plates.

The detection of groundwater flow by precise temperature measurements in boreholes

The flow of water is a very effective means for the transfer of heat, and one method of detecting such flow is to make precise temperature measurements at closely spaced intervals in a borehole that intersects a flow zone. Water can flow through permeable formations; within a borehole it can flow between two aquifers or fracture systems; it can flow into a fracture system during the drilling of a borehole; and it can flow up or down narrow, dipping fracture zones. Each of these phenomena produces a characteristic thermal signature on a borehole temperature log that can be modelled mathematically. Analysis of such thermal anomalies permits, therefore, a quantitative estimate to be made of the amount and rate of fluid flow. In principle, very small flow rates can be detected from their thermal effects, but in practice other factors, such as thermal conductivity variations, can cause variations in thermal gradients that limit the detectability. Anomalies that persist over large depth ranges compared with the diameter of the borehole can generally be interpreted unambiguously. Examples of each type of flow are given.

Heat flow and heat generation in the Churchill Province of the Canadian Shield, and their palaeotectonic significance☆

Abstract Six new heat flow determinations are presented for Proterozoic mobile belts of the Churchill Province of the Canadian Shield, an area that was affected by several stages of the Hudsonian orogenic sequence (1.9-1.6 Ga ago). With other, previously published, values the mean of eight determinations considered reliable and representative and corrected for the effects of Pleistocene glaciation is 44 ± 7 mW m−2. Heat generation measurements have also been made; values range from 0.1–1.04 μW m−3. A linear relation between heat flow and heat production is apparent. The heat flow axis intercept is 37 mW m−2, and the scale depth is 11 km, compared with 28 mW m−2 and 13.6 km for the Archaean Superior Province. Approximately 20% of the Churchill heat flow appears to be derived from radioactive decay in the upper crust, compared with 30% for the Superior Province and shields as a whole. The observations imply that the heat flow-heat production relation for the Churchill Province should be written as Q = Qc + Qe + A0b where Qc is equivalent to the reduced heat flow for the Archaean terrain, b is similar for the two, and Qe is an additional component of heat flow in the Proterozoic mobile belts of the Churchill Province. A speculative tectonic model is presented. It is suggested that rifting along two axes of an original craton, which had lateral variations in near surface radiogenic element concentration, followed by erosion of the radiogenic layer and subsequent reconvergence of the cratonic segments, led to widespread redistribution of radioactive elements into the reactivated inter-rift crustal block. One result would be that crustal temperatures are higher in that part of the Churchill Province than in the Superior.

Water movement within lac du bonnet batholith as revealed by detailed thermal studies of three closely-spaced boreholes

Abstract Successive temperature logs have been obtained over a period of two years in three closely-spaced boreholes in the Lac du Bonnet batholith of the Superior Province of the Canadian Shield. Two of the boreholes, of depth 450 m and 830 m, intersect a dipping fracture zone at 435–450 m. In both holes water is flowing from near the surface to the fracture zone at approximately 1.5–1.9·10−5 m3 s−1, the flow being inferred from analysis of the temperature logs. Below 25 m, temperatures in these two holes are 0.22–0.28 K lower than those in the third, 145 m, hole. The temperature data have been combined with over 200 thermal conductivity measurements on core samples to produce heat flow values. In the deepest hole heat flow above the fracture zone is 16% higher than that below the zone. This indicates that water is flowing up the fracture zone. The flow rate is approximately 0.3 g s−1 m−1, and the flow has existed for thousands of years. Observation of thermal effects of water flow in massive, relatively unfractured plutons in a region having little topographic relief causes one to be concerned about the reliability of heat flow values measured in similar environments. The regional heat flow is taken to be 50 mW m−2 after correction for glaciation effects. The average value of 24 determinations of radioactive heat generation in granitic core samples is 5.23 ± 1.11 μW m−3, which is more than three times higher than expected for such a heat flow in the Superior Province. This implies that the layer of high radioactive heat generation is thin, being not more than 4 km and probably about 1.3 km thick.

Heat flow provinces reconsidered

Abstract The concept of a heat flow province, an area in which there is apparently a linear relationship between near-surface heat flow and radiogenic heat generation, is considered. The relationship has been suggested for a wide variety of tectonic terrains, ranging in age from Archaean to Cenozoic. It is expressed as q = q 0 + A 0 D in which q and A o are the measured heat flow and heat generation and q o and D are interpreted to be the uniform reduced heat flow from the mantle and lower crust, and D is a scale depth for the distribution of radiogenic elements. Reported values for D are in a remarkably narrow range, and are typically ∼ 10 km. However, there can be substantial uncertainties associated with the data. In particular, heat generation measurements from a limited number of sites may not adequately represent even a relatively uniform pluton. When the data are re-analysed with reasonable estimates of uncertainty associated with them, the calculated values of q o and D are, in most cases, significantly different from those calculated without regard to uncertainties in the data. Only the data from the Sierra Nevada batholith, a deep-seated plutonic complex, appear to satisfy the relationship even when their uncertainties are taken into account. With the re-analysis, D generally falls in the approximate range 10–20 km. This is also the range of depths to a commonly-observed discontinuity in the mid- to lower-crust, sometimes called the Conrad discontinuity, which suggests a possible correlation between the two. A model in which blocks of crust 10–20 km thick, depths being randomly arranged, are each assigned a uniform heat generation, randomly generated from a reasonable range, produces an apparent relationship between heat flow and heat generation that is at least as good as most of those deduced from real data. The model produces crustal temperatures that are generally considerably lower in the lower crust than those commonly calculated; this is consistent with the observations from seismic refraction that the depth to the Moho can vary widely even in stable shield environments. It is concluded that the concept of a heat flow province as originally expounded is invalid.

The thermal nature of the Canadian Appalachian crust

Abstract Heat flow values for 17 new sites in the Canadian Appalachians of Nova Scotia, Prince Edward Island, New Brunswick and Quebec are reported. They consist of ten high-quality measurements obtained from a combination of accurate temperature gradient and thermal conductivity measurements, and seven values obtained from sites at which accurate temperature data were obtained but at which lack of core material meant that conductivity had to be estimated from lithological information. The mean and standard deviation of heat flow from the new sites is 58 ± 11 mW/m 2 , not significantly different from the mean and standard deviation obtained by incorporating seventeen previously published values, 57 ± 11 mW/m 2 . Heat flow varies across this part of the Canadian Appalachians, being lowest in the central part, which is underlain by a major Carboniferous basin, and highest near the coast. There is no obvious relationship between heat flow and age of orogenic imprint. In southern New Brunswick there is an area of high heat flow, greater than 70 mW/m 2 , that appears to be related to the presence of radiogenic Devonian granitic batholiths, both exposed and buried. Using a previously defined heat flow-heat generation relationship for the Appalachians of Canada, the radiogenic layer is estimated to be approximately 1.4 to 3.3 km thick in the St. George batholith of New Brunswick, and surficial in the Wedgeport pluton of Nova Scotia. Gravity interpretation suggests that the maximum thickness of the St. George batholith is 7.5 km. Very low values of heat flow are reported for the Magdalen Basin. Up to 10 km of crust may have been eroded from the basement prior to the formation of the southern part of the basin, removing much of the radiogenic source of heat.

Perturbations to temperature gradients by water flow in crystalline rock formations

Abstract The flow of water can provide a very effective means for the transfer of heat. Closely-spaced (≈- 3 m) temperature measurements in boreholes drilled into crystalline rock bodies have revealed the thermal effects of water flow that in extreme cases seriously disturb the purely conductive temperature gradient. Such flows appear to be widespread in and between fractures in what are otherwise impermeable rock bodies. In the typical heat-flux determination, the vertical and horizontal spacing of temperature measurements might be insufficient to indicate localised water flows. Small flows can, however, produce a significant error in measurement of conductive heat-flux. It is recommended that boreholes should be logged at no more than 5 m intervals over their entire length, and that ideally several holes, with a horizontal separation similar to the average hole depth, should be logged, in order that the thermal effects of water flow can be recognized.

The heat flow—heat generation relationship: Implications for the nature of continental crust

Abstract The apparent linear relationship between surface heat flow, Q and heat generation, A, is usually interpreted in terms of heat flow provinces i.e., terrains that have a common tectonothermal history and a uniform mantle component of heat flow. The slope, D, of the linear regression of heat flow on heat generation is an indication of the thickness of the upper heat-producing layer. For many terrains D is numerically similar to the depth of a major mid-crustal transition zone. These observations are modelled by a two-layer crust consisting of an upper layer, approximately 10–15 km thick, of variable heat generation, underlain by a layer about 20 km thick in which heat generation is lower and more uniform than that in the upper layer. The model is well able to simulate Q-A regression parameters for different sets of real data. It is also consistent with other geological and geophysical observations. The higher and more variable heat generation of the upper layer is consistent with ideas concerning removal of heat-producing elements from the lower layer during amphibolite to granulite metamorphism and their redistribution into the upper crust. Further redistribution of heat-producing elements may occur in zones of hydrothermal circulation whose depth is a function of stress, pressure and temperature, and which may migrate vertically in response to stress changes during uplift and erosion. Large-scale metamorphism to granulite facies has been postulated as one mechanism for stabilisation of continental crust in the late Archaean. It is shown, however, that redistribution of heat-producing elements in such a process would have only a small effect on crustal geotherms, so that partitioning of heat production was not a major factor in the crustal stabilisation.

Thermal diffusivity of some crystalline rocks

Abstract Thermal diffusivity data at room temperature and uniaxial pressure of 1 MPa are reported for five sets of crystalline rocks—granite, granodiorite, gabbro, basalt and gneiss. Diffusivity ranges between approximately 0.6 and 1.9 mm 2 /s, the lower end of the range being appropriate for basic rocks and the upper end for quartz-bearing acidic rocks. The scatter in diffusivity for each data set is significantly more than that of thermal conductivity, because the diffusivity of water is typically less than 10% of the diffusivity of most common minerals, whereas water conductivity is 25–30% of the conductivity of the minerals. For a sample set of uniform mineralogy in which porosity varies, a greater variation of diffusivity than of conductivity is therefore expected. For three of the sets sufficient mineralogical data were available to permit the assessment of methods of estimating thermal diffusivity from mineral content. All models tested yielded higher mean values of diffusivity than the means of the measured values. No model was found to be able to predict diffusivity to better than approximately 20%, but if that accuracy is sufficient, a simple geometrical model, for which only quartz content must be known, is adequate. The diffusivity data have been combined with measurements of thermal conductivity and density to provide estimates of specific heat. These all tend to be higher than those reported in the literature. For some rocks, such as the basalts, this can be explained in terms of relatively high water content and the very high specific heat of water compared with that of most common minerals. For the granites and granodiorites, the new specific heat data redefine the previously published means and ranges, by increasing the data base by approximately an order of magnitude.

Tectonothermics of the North American great plains basement

Abstract Heat flow and heat generation data from the Superior, Churchill and Wyoming Provinces of western North America contribute to our understanding of the nature of the crust beneath the sedimentary cover of the Great Plains. A major structural boundary defined by gravity gradient data (Thomas et al., 1987) splits the terrains between the cratons into two distinct geothermal provinces. The Wyoming and Churchill Provinces are characterised by higher heat flow than the Superior Province and the eastern part of the intervening Hudsonian terrain, the Proterozoic Mobile Belt (PMB). Heat flow in the PMB is typical of that of the Superior craton. This suggests, on the basis of a model by Ballard and Pollack (1987), that the PMB is underlain by Archaean cratonic lithosphere. The similarity in heat flow between the PMB and the Superior craton suggests that the crust of the PMB is akin to that of the Superior craton rather than to that of the more radiogenic Wyoming and Churchill cratons. Geochronologic, gravity, aeromagnetic and geothermal data suggest that the structural boundary defining the western boundary of the PMB is also the eastern boundary of the Wyoming craton. Geothermal data alone suggest that the Wyoming craton might extend as far north as the cratonic segment of the Churchill Province, or that the Wyoming and Churchill Provinces may gradually merge together.